Project Candle bug triage (18th October 2015)

[Nicholas Nethercote]

Greetings,

This is Project Candle bug triage email. The process used is described at
https://wiki.mozilla.org/Performance/Project_Candle#Bug_triage_process.

At the time of writing there are 20 unprioritized Power bugs, viewable at
http://mzl.la/1GdpKBp. In this thread I am discussing 10 of them.


* https://bugzilla.mozilla.org/show_bug.cgi?id=834343
* nsIdleService reports active topic even when there is no activity

Not sure about this one. It could be totally bogus. Any ideas?


* https://bugzilla.mozilla.org/show_bug.cgi?id=845522
* Slow rendering: loop of « Freeing buffer » & « Allocated buffer » may affect
power usage

Probably no longer relevant, though these sentences appear to be contradictory:
"I have not tried to reproduce it since. I cannot reproduce this issue with
the STR provided in comment 0." I've needinfo'd npb.


* https://bugzilla.mozilla.org/show_bug.cgi?id=890154
* [system] Power consumption increases when using Facebook application.

Judging from comment 37 this bug may be closeable. I've asked in the bug.


* https://bugzilla.mozilla.org/show_bug.cgi?id=895991
* Background services should stop when we receive battery and network warning
intents

I suggest P2 in the absence of any data about how much effect this will have.
mfinkle, rnewman: any suggestions?


* https://bugzilla.mozilla.org/show_bug.cgi?id=915663
* [META] Power draw testing setup work.
and
* https://bugzilla.mozilla.org/show_bug.cgi?id=915719
* Baseline power draw measurements of Android ICS and/or JB on Nexus 4

The former is a meta-bug. The latter is one of the two open bugs blocking it.
The other bug blocking is is bug 915725 which is a Power:P1.

Perhaps the former could be Power:meta and the latter could be... P2? I kind of
feel like the latter could be dup'd to 915725, since that bug sounds like it
encompasses 915719. jhylands, any thoughts?


* https://bugzilla.mozilla.org/show_bug.cgi?id=927377
* Huge CPU use on http://frozenbyte.com/ with image high-quality downscaling
enabled
and
* https://bugzilla.mozilla.org/show_bug.cgi?id=953037
* 100% CPU Load, no JS, CSS only. Bootstrap 3.0.3. Probably faulty page
rendering engine

Both involve high-quality downscaling. tn thinks this might be fixed. I've
asked Loic and Alice if they can still reproduce. I suggest P2 if so.


* https://bugzilla.mozilla.org/show_bug.cgi?id=974207
* SafeBrowsing does frequent network activity

"Frequent" here is once every 45 minutes, which doesn't seem that bad to me.
P3? WONTFIX? mfinkle, what do you think?


* https://bugzilla.mozilla.org/show_bug.cgi?id=1214208
* Power Consumption increased in music app while playing music

A fresh B2G power regression. P2? jhylands, gsvelto: what do you think?

[Kan-Ru Chen (陳侃如)]

Nicholas Nethercote <n.neth...@gmail.com> writes:

> * https://bugzilla.mozilla.org/show_bug.cgi?id=834343
> * nsIdleService reports active topic even when there is no activity
>
> Not sure about this one. It could be totally bogus. Any ideas?

Replied to the bug. Maybe P3 or WONTFIX?

Kanru

[Gabriele Svelto]

On 19/10/2015 01:53, Nicholas Nethercote wrote:
>
> * https://bugzilla.mozilla.org/show_bug.cgi?id=1214208
> * Power Consumption increased in music app while playing music
>
> A fresh B2G power regression. P2? jhylands, gsvelto: what do you think?

Sounds bad as playing MP3s is a very common use case. Also in the past
our partners have compared this scenario's battery life against Android
running on the same hardware and rejected builds if our consumption was
measurably higher (>10%). I'd say P1 because of this and in fact I'd
even ask for it to block the 2.5 release; IMHO we can't ship with a
power regression on such a common use-case.

Gabriele

[Jonathan Hylands]

On Sun, Oct 18, 2015 at 7:53 PM, Nicholas Nethercote <n.neth...@gmail.com> wrote:

Greetings,

* https://bugzilla.mozilla.org/show_bug.cgi?id=915663
* [META] Power draw testing setup work.
and
* https://bugzilla.mozilla.org/show_bug.cgi?id=915719
* Baseline power draw measurements of Android ICS and/or JB on Nexus 4

The former is a meta-bug. The latter is one of the two open bugs blocking it.
The other bug blocking is is bug 915725 which is a Power:P1.

Perhaps the former could be Power:meta and the latter could be... P2? I kind of
feel like the latter could be dup'd to 915725, since that bug sounds like it
encompasses 915719. jhylands, any thoughts?

I have on my Q4 list of deliverables a task to set up a new comparative baseline between Android and FxOS on Aries, so this is definitely relevant again, although I think we should close both of these and I'll create a new bug when I'm working on that piece.

915725 is another of the same, although for browsers under Android, and perhaps with the new Galaxy S4 harness we can get someone under mfinkle to look at doing that.

 

* https://bugzilla.mozilla.org/show_bug.cgi?id=1214208
* Power Consumption increased in music app while playing music

A fresh B2G power regression. P2? jhylands, gsvelto: what do you think?

I agree with gsvelto - this is important, and needs to be fixed ASAP. I bisected it last week and determined which commit caused it, which is documented in the bug.

[Nicholas Nethercote]

Thank you to all those who participated in the discussion.

On Sun, Oct 18, 2015 at 4:53 PM, Nicholas Nethercote
<n.neth...@gmail.com> wrote:
>
> * https://bugzilla.mozilla.org/show_bug.cgi?id=834343
> * nsIdleService reports active topic even when there is no activity
>
> Not sure about this one. It could be totally bogus. Any ideas?

P3 on Kan-Ru's suggestion.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=845522
> * Slow rendering: loop of « Freeing buffer » & « Allocated buffer » may affect
> power usage
>
> Probably no longer relevant, though these sentences appear to be contradictory:
> "I have not tried to reproduce it since. I cannot reproduce this issue with
> the STR provided in comment 0." I've needinfo'd npb.

It got resolved as WORKSFORME during the week.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=890154
> * [system] Power consumption increases when using Facebook application.
>
> Judging from comment 37 this bug may be closeable. I've asked in the bug.

It got resolved as FIXED during the week.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=895991
> * Background services should stop when we receive battery and network warning
> intents
>
> I suggest P2 in the absence of any data about how much effect this will have.
> mfinkle, rnewman: any suggestions?

P2 in the absence of other suggestions.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=915663
> * [META] Power draw testing setup work.
> and
> * https://bugzilla.mozilla.org/show_bug.cgi?id=915719
> * Baseline power draw measurements of Android ICS and/or JB on Nexus 4
>
> The former is a meta-bug. The latter is one of the two open bugs blocking it.
> The other bug blocking is is bug 915725 which is a Power:P1.
>
> Perhaps the former could be Power:meta and the latter could be... P2? I kind of
> feel like the latter could be dup'd to 915725, since that bug sounds like it
> encompasses 915719. jhylands, any thoughts?

I closed both of these (INCOMPLETE) on jhylands' suggestion. He'll
open new bugs when appropriate.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=927377
> * Huge CPU use on http://frozenbyte.com/ with image high-quality downscaling
> enabled
> and
> * https://bugzilla.mozilla.org/show_bug.cgi?id=953037
> * 100% CPU Load, no JS, CSS only. Bootstrap 3.0.3. Probably faulty page
> rendering engine
>
> Both involve high-quality downscaling. tn thinks this might be fixed. I've
> asked Loic and Alice if they can still reproduce. I suggest P2 if so.

Both were resolved as WORKSFORME during the week because high-quality
downscaling has been removed in favour of downscale-during-decode.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=974207
> * SafeBrowsing does frequent network activity
>
> "Frequent" here is once every 45 minutes, which doesn't seem that bad to me.
> P3? WONTFIX? mfinkle, what do you think?

P3 in absence of other suggestions.

> * https://bugzilla.mozilla.org/show_bug.cgi?id=1214208
> * Power Consumption increased in music app while playing music
>
> A fresh B2G power regression. P2? jhylands, gsvelto: what do you think?

P1 on gsvelto's suggestion.